%matplotlib inline
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np
from sklearn.model_selection import train_test_split
from subprocess import check_output
from sklearn.metrics import confusion_matrix
import numpy as np
from sklearn.manifold import TSNE
from sklearn.model_selection import GridSearchCV
train = pd.read_csv("pml-training.csv", error_bad_lines=False, index_col=False, dtype='unicode')
train.head()
| Unnamed: 0 | user_name | raw_timestamp_part_1 | raw_timestamp_part_2 | cvtd_timestamp | new_window | num_window | roll_belt | pitch_belt | yaw_belt | ... | gyros_forearm_x | gyros_forearm_y | gyros_forearm_z | accel_forearm_x | accel_forearm_y | accel_forearm_z | magnet_forearm_x | magnet_forearm_y | magnet_forearm_z | classe | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | carlitos | 1323084231 | 788290 | 05/12/2011 11:23 | no | 11 | 1.41 | 8.07 | -94.4 | ... | 0.03 | 0 | -0.02 | 192 | 203 | -215 | -17 | 654 | 476 | A |
| 1 | 2 | carlitos | 1323084231 | 808298 | 05/12/2011 11:23 | no | 11 | 1.41 | 8.07 | -94.4 | ... | 0.02 | 0 | -0.02 | 192 | 203 | -216 | -18 | 661 | 473 | A |
| 2 | 3 | carlitos | 1323084231 | 820366 | 05/12/2011 11:23 | no | 11 | 1.42 | 8.07 | -94.4 | ... | 0.03 | -0.02 | 0 | 196 | 204 | -213 | -18 | 658 | 469 | A |
| 3 | 4 | carlitos | 1323084232 | 120339 | 05/12/2011 11:23 | no | 12 | 1.48 | 8.05 | -94.4 | ... | 0.02 | -0.02 | 0 | 189 | 206 | -214 | -16 | 658 | 469 | A |
| 4 | 5 | carlitos | 1323084232 | 196328 | 05/12/2011 11:23 | no | 12 | 1.48 | 8.07 | -94.4 | ... | 0.02 | 0 | -0.02 | 189 | 206 | -214 | -17 | 655 | 473 | A |
5 rows × 160 columns
train.dropna(axis=1,thresh=int(0.20*train.shape[0]),inplace=True)
train.isnull().sum()
Unnamed: 0 0 user_name 0 raw_timestamp_part_1 0 raw_timestamp_part_2 0 cvtd_timestamp 0 new_window 0 num_window 0 roll_belt 0 pitch_belt 0 yaw_belt 0 total_accel_belt 0 gyros_belt_x 0 gyros_belt_y 0 gyros_belt_z 0 accel_belt_x 0 accel_belt_y 0 accel_belt_z 0 magnet_belt_x 0 magnet_belt_y 0 magnet_belt_z 0 roll_arm 0 pitch_arm 0 yaw_arm 0 total_accel_arm 0 gyros_arm_x 0 gyros_arm_y 0 gyros_arm_z 0 accel_arm_x 0 accel_arm_y 0 accel_arm_z 0 magnet_arm_x 0 magnet_arm_y 0 magnet_arm_z 0 roll_dumbbell 0 pitch_dumbbell 0 yaw_dumbbell 0 total_accel_dumbbell 0 gyros_dumbbell_x 0 gyros_dumbbell_y 0 gyros_dumbbell_z 0 accel_dumbbell_x 0 accel_dumbbell_y 0 accel_dumbbell_z 0 magnet_dumbbell_x 0 magnet_dumbbell_y 0 magnet_dumbbell_z 0 roll_forearm 0 pitch_forearm 0 yaw_forearm 0 total_accel_forearm 0 gyros_forearm_x 0 gyros_forearm_y 0 gyros_forearm_z 0 accel_forearm_x 0 accel_forearm_y 0 accel_forearm_z 0 magnet_forearm_x 0 magnet_forearm_y 0 magnet_forearm_z 0 classe 0 dtype: int64
train.columns
Index(['Unnamed: 0', 'user_name', 'raw_timestamp_part_1',
'raw_timestamp_part_2', 'cvtd_timestamp', 'new_window', 'num_window',
'roll_belt', 'pitch_belt', 'yaw_belt', 'total_accel_belt',
'gyros_belt_x', 'gyros_belt_y', 'gyros_belt_z', 'accel_belt_x',
'accel_belt_y', 'accel_belt_z', 'magnet_belt_x', 'magnet_belt_y',
'magnet_belt_z', 'roll_arm', 'pitch_arm', 'yaw_arm', 'total_accel_arm',
'gyros_arm_x', 'gyros_arm_y', 'gyros_arm_z', 'accel_arm_x',
'accel_arm_y', 'accel_arm_z', 'magnet_arm_x', 'magnet_arm_y',
'magnet_arm_z', 'roll_dumbbell', 'pitch_dumbbell', 'yaw_dumbbell',
'total_accel_dumbbell', 'gyros_dumbbell_x', 'gyros_dumbbell_y',
'gyros_dumbbell_z', 'accel_dumbbell_x', 'accel_dumbbell_y',
'accel_dumbbell_z', 'magnet_dumbbell_x', 'magnet_dumbbell_y',
'magnet_dumbbell_z', 'roll_forearm', 'pitch_forearm', 'yaw_forearm',
'total_accel_forearm', 'gyros_forearm_x', 'gyros_forearm_y',
'gyros_forearm_z', 'accel_forearm_x', 'accel_forearm_y',
'accel_forearm_z', 'magnet_forearm_x', 'magnet_forearm_y',
'magnet_forearm_z', 'classe'],
dtype='object')
sns.factorplot(x="classe", data=train, kind="count", palette="Set2")
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/categorical.py:3704: UserWarning: The `factorplot` function has been renamed to `catplot`. The original name will be removed in a future release. Please update your code. Note that the default `kind` in `factorplot` (`'point'`) has changed `'strip'` in `catplot`. warnings.warn(msg)
<seaborn.axisgrid.FacetGrid at 0x7f95c8f912e0>
def perform_tsne(X_data, y_data, perplexities, n_iter=1000, img_name_prefix='t-sne'):
for index,perplexity in enumerate(perplexities):
# perform t-sne
print('\nperforming tsne with perplexity {} and with {} iterations at max'.format(perplexity, n_iter))
X_reduced = TSNE(verbose=2, perplexity=perplexity).fit_transform(X_data)
print('Done..')
# prepare the data for seaborn
print('Creating plot for this t-sne visualization..')
df = pd.DataFrame({'x':X_reduced[:,0], 'y':X_reduced[:,1] ,'label':y_data})
# draw the plot in appropriate place in the grid
sns.lmplot(data=df, x='x', y='y', hue='label', fit_reg=False, size=8,\
palette="Set1",markers=[1,2,3,4,5])
plt.title("perplexity : {} and max_iter : {}".format(perplexity, n_iter))
img_name = img_name_prefix + '_perp_{}_iter_{}.png'.format(perplexity, n_iter)
print('saving this plot as image in present working directory...')
plt.savefig(img_name)
plt.show()
print('Done')
train['new_window'] = train['new_window'].map( {'no': 0, 'yes': 1} ).astype(int)
train = train.drop(['Unnamed: 0','cvtd_timestamp'], axis=1)
train['user_name'] = train['user_name'].map( {'carlitos': 0, 'pedro': 1, 'adelmo': 2, 'charles': 3,
'eurico': 4, 'jeremy': 5} ).astype(int)
X_pre_tsne = train.drop(['classe'], axis=1)
y_pre_tsne = train['classe']
perform_tsne(X_data = X_pre_tsne,y_data=y_pre_tsne, perplexities =[2,10,20,50])
performing tsne with perplexity 2 and with 1000 iterations at max [t-SNE] Computing 7 nearest neighbors... [t-SNE] Indexed 19622 samples in 0.046s... [t-SNE] Computed neighbors for 19622 samples in 0.412s... [t-SNE] Computed conditional probabilities for sample 1000 / 19622 [t-SNE] Computed conditional probabilities for sample 2000 / 19622 [t-SNE] Computed conditional probabilities for sample 3000 / 19622 [t-SNE] Computed conditional probabilities for sample 4000 / 19622 [t-SNE] Computed conditional probabilities for sample 5000 / 19622 [t-SNE] Computed conditional probabilities for sample 6000 / 19622 [t-SNE] Computed conditional probabilities for sample 7000 / 19622 [t-SNE] Computed conditional probabilities for sample 8000 / 19622 [t-SNE] Computed conditional probabilities for sample 9000 / 19622 [t-SNE] Computed conditional probabilities for sample 10000 / 19622 [t-SNE] Computed conditional probabilities for sample 11000 / 19622 [t-SNE] Computed conditional probabilities for sample 12000 / 19622 [t-SNE] Computed conditional probabilities for sample 13000 / 19622 [t-SNE] Computed conditional probabilities for sample 14000 / 19622 [t-SNE] Computed conditional probabilities for sample 15000 / 19622 [t-SNE] Computed conditional probabilities for sample 16000 / 19622 [t-SNE] Computed conditional probabilities for sample 17000 / 19622 [t-SNE] Computed conditional probabilities for sample 18000 / 19622 [t-SNE] Computed conditional probabilities for sample 19000 / 19622 [t-SNE] Computed conditional probabilities for sample 19622 / 19622 [t-SNE] Mean sigma: 87.467530 [t-SNE] Computed conditional probabilities in 0.089s [t-SNE] Iteration 50: error = 136.9432983, gradient norm = 0.0128047 (50 iterations in 4.016s) [t-SNE] Iteration 100: error = 115.7826691, gradient norm = 0.0076069 (50 iterations in 3.512s) [t-SNE] Iteration 150: error = 108.7137909, gradient norm = 0.0053407 (50 iterations in 3.546s) [t-SNE] Iteration 200: error = 104.5358276, gradient norm = 0.0044148 (50 iterations in 3.692s) [t-SNE] Iteration 250: error = 101.5662231, gradient norm = 0.0038067 (50 iterations in 3.127s) [t-SNE] KL divergence after 250 iterations with early exaggeration: 101.566223 [t-SNE] Iteration 300: error = 5.4225655, gradient norm = 0.0012851 (50 iterations in 3.593s) [t-SNE] Iteration 350: error = 4.6302996, gradient norm = 0.0008195 (50 iterations in 3.649s) [t-SNE] Iteration 400: error = 4.0268984, gradient norm = 0.0005793 (50 iterations in 3.620s) [t-SNE] Iteration 450: error = 3.5745971, gradient norm = 0.0004507 (50 iterations in 3.728s) [t-SNE] Iteration 500: error = 3.2329259, gradient norm = 0.0003902 (50 iterations in 3.521s) [t-SNE] Iteration 550: error = 2.9854610, gradient norm = 0.0003263 (50 iterations in 3.509s) [t-SNE] Iteration 600: error = 2.7919118, gradient norm = 0.0002901 (50 iterations in 3.522s) [t-SNE] Iteration 650: error = 2.6328263, gradient norm = 0.0002634 (50 iterations in 3.546s) [t-SNE] Iteration 700: error = 2.5015061, gradient norm = 0.0002405 (50 iterations in 3.614s) [t-SNE] Iteration 750: error = 2.3898180, gradient norm = 0.0002208 (50 iterations in 3.592s) [t-SNE] Iteration 800: error = 2.2931180, gradient norm = 0.0002054 (50 iterations in 3.618s) [t-SNE] Iteration 850: error = 2.2082231, gradient norm = 0.0001918 (50 iterations in 3.543s) [t-SNE] Iteration 900: error = 2.1328387, gradient norm = 0.0001808 (50 iterations in 3.500s) [t-SNE] Iteration 950: error = 2.0652199, gradient norm = 0.0001707 (50 iterations in 3.505s) [t-SNE] Iteration 1000: error = 2.0040934, gradient norm = 0.0001615 (50 iterations in 3.510s) [t-SNE] KL divergence after 1000 iterations: 2.004093 Done.. Creating plot for this t-sne visualization..
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/regression.py:580: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
saving this plot as image in present working directory...
Done performing tsne with perplexity 10 and with 1000 iterations at max [t-SNE] Computing 31 nearest neighbors... [t-SNE] Indexed 19622 samples in 0.048s... [t-SNE] Computed neighbors for 19622 samples in 0.604s... [t-SNE] Computed conditional probabilities for sample 1000 / 19622 [t-SNE] Computed conditional probabilities for sample 2000 / 19622 [t-SNE] Computed conditional probabilities for sample 3000 / 19622 [t-SNE] Computed conditional probabilities for sample 4000 / 19622 [t-SNE] Computed conditional probabilities for sample 5000 / 19622 [t-SNE] Computed conditional probabilities for sample 6000 / 19622 [t-SNE] Computed conditional probabilities for sample 7000 / 19622 [t-SNE] Computed conditional probabilities for sample 8000 / 19622 [t-SNE] Computed conditional probabilities for sample 9000 / 19622 [t-SNE] Computed conditional probabilities for sample 10000 / 19622 [t-SNE] Computed conditional probabilities for sample 11000 / 19622 [t-SNE] Computed conditional probabilities for sample 12000 / 19622 [t-SNE] Computed conditional probabilities for sample 13000 / 19622 [t-SNE] Computed conditional probabilities for sample 14000 / 19622 [t-SNE] Computed conditional probabilities for sample 15000 / 19622 [t-SNE] Computed conditional probabilities for sample 16000 / 19622 [t-SNE] Computed conditional probabilities for sample 17000 / 19622 [t-SNE] Computed conditional probabilities for sample 18000 / 19622 [t-SNE] Computed conditional probabilities for sample 19000 / 19622 [t-SNE] Computed conditional probabilities for sample 19622 / 19622 [t-SNE] Mean sigma: 123.379298 [t-SNE] Computed conditional probabilities in 0.329s [t-SNE] Iteration 50: error = 113.7130890, gradient norm = 0.0369656 (50 iterations in 3.634s) [t-SNE] Iteration 100: error = 94.1610870, gradient norm = 0.0043113 (50 iterations in 3.262s) [t-SNE] Iteration 150: error = 87.3512268, gradient norm = 0.0031370 (50 iterations in 3.152s) [t-SNE] Iteration 200: error = 83.2583008, gradient norm = 0.0026087 (50 iterations in 3.139s) [t-SNE] Iteration 250: error = 80.3271103, gradient norm = 0.0022567 (50 iterations in 3.092s) [t-SNE] KL divergence after 250 iterations with early exaggeration: 80.327110 [t-SNE] Iteration 300: error = 3.7938137, gradient norm = 0.0011539 (50 iterations in 3.397s) [t-SNE] Iteration 350: error = 3.1151946, gradient norm = 0.0007770 (50 iterations in 3.868s) [t-SNE] Iteration 400: error = 2.5685015, gradient norm = 0.0005510 (50 iterations in 3.826s) [t-SNE] Iteration 450: error = 2.1605334, gradient norm = 0.0004132 (50 iterations in 3.878s) [t-SNE] Iteration 500: error = 1.8564134, gradient norm = 0.0003219 (50 iterations in 3.926s) [t-SNE] Iteration 550: error = 1.6248825, gradient norm = 0.0002599 (50 iterations in 3.910s) [t-SNE] Iteration 600: error = 1.4440322, gradient norm = 0.0002150 (50 iterations in 3.828s) [t-SNE] Iteration 650: error = 1.2993627, gradient norm = 0.0001818 (50 iterations in 3.696s) [t-SNE] Iteration 700: error = 1.1817625, gradient norm = 0.0001565 (50 iterations in 3.515s) [t-SNE] Iteration 750: error = 1.0844851, gradient norm = 0.0001381 (50 iterations in 3.534s) [t-SNE] Iteration 800: error = 1.0035207, gradient norm = 0.0001227 (50 iterations in 3.955s) [t-SNE] Iteration 850: error = 0.9356239, gradient norm = 0.0001105 (50 iterations in 3.484s) [t-SNE] Iteration 900: error = 0.8784217, gradient norm = 0.0001003 (50 iterations in 3.353s) [t-SNE] Iteration 950: error = 0.8295435, gradient norm = 0.0000923 (50 iterations in 3.331s) [t-SNE] Iteration 1000: error = 0.7871184, gradient norm = 0.0000848 (50 iterations in 3.318s) [t-SNE] KL divergence after 1000 iterations: 0.787118 Done.. Creating plot for this t-sne visualization.. saving this plot as image in present working directory...
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/regression.py:580: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
Done performing tsne with perplexity 20 and with 1000 iterations at max [t-SNE] Computing 61 nearest neighbors... [t-SNE] Indexed 19622 samples in 0.051s... [t-SNE] Computed neighbors for 19622 samples in 0.815s... [t-SNE] Computed conditional probabilities for sample 1000 / 19622 [t-SNE] Computed conditional probabilities for sample 2000 / 19622 [t-SNE] Computed conditional probabilities for sample 3000 / 19622 [t-SNE] Computed conditional probabilities for sample 4000 / 19622 [t-SNE] Computed conditional probabilities for sample 5000 / 19622 [t-SNE] Computed conditional probabilities for sample 6000 / 19622 [t-SNE] Computed conditional probabilities for sample 7000 / 19622 [t-SNE] Computed conditional probabilities for sample 8000 / 19622 [t-SNE] Computed conditional probabilities for sample 9000 / 19622 [t-SNE] Computed conditional probabilities for sample 10000 / 19622 [t-SNE] Computed conditional probabilities for sample 11000 / 19622 [t-SNE] Computed conditional probabilities for sample 12000 / 19622 [t-SNE] Computed conditional probabilities for sample 13000 / 19622 [t-SNE] Computed conditional probabilities for sample 14000 / 19622 [t-SNE] Computed conditional probabilities for sample 15000 / 19622 [t-SNE] Computed conditional probabilities for sample 16000 / 19622 [t-SNE] Computed conditional probabilities for sample 17000 / 19622 [t-SNE] Computed conditional probabilities for sample 18000 / 19622 [t-SNE] Computed conditional probabilities for sample 19000 / 19622 [t-SNE] Computed conditional probabilities for sample 19622 / 19622 [t-SNE] Mean sigma: 137.904129 [t-SNE] Computed conditional probabilities in 0.571s [t-SNE] Iteration 50: error = 109.5277710, gradient norm = 0.0113789 (50 iterations in 3.749s) [t-SNE] Iteration 100: error = 87.3699951, gradient norm = 0.0037697 (50 iterations in 3.392s) [t-SNE] Iteration 150: error = 80.7580490, gradient norm = 0.0027438 (50 iterations in 3.296s) [t-SNE] Iteration 200: error = 76.7914734, gradient norm = 0.0022827 (50 iterations in 3.246s) [t-SNE] Iteration 250: error = 73.9716949, gradient norm = 0.0020191 (50 iterations in 3.367s) [t-SNE] KL divergence after 250 iterations with early exaggeration: 73.971695 [t-SNE] Iteration 300: error = 3.2370901, gradient norm = 0.0011675 (50 iterations in 3.373s) [t-SNE] Iteration 350: error = 2.5716629, gradient norm = 0.0007569 (50 iterations in 3.384s) [t-SNE] Iteration 400: error = 2.0657315, gradient norm = 0.0005242 (50 iterations in 3.459s) [t-SNE] Iteration 450: error = 1.7014210, gradient norm = 0.0003873 (50 iterations in 3.578s) [t-SNE] Iteration 500: error = 1.4347816, gradient norm = 0.0002992 (50 iterations in 3.936s) [t-SNE] Iteration 550: error = 1.2344266, gradient norm = 0.0002392 (50 iterations in 4.030s) [t-SNE] Iteration 600: error = 1.0808647, gradient norm = 0.0001961 (50 iterations in 4.111s) [t-SNE] Iteration 650: error = 0.9607904, gradient norm = 0.0001642 (50 iterations in 3.861s) [t-SNE] Iteration 700: error = 0.8651028, gradient norm = 0.0001397 (50 iterations in 3.875s) [t-SNE] Iteration 750: error = 0.7876707, gradient norm = 0.0001209 (50 iterations in 3.916s) [t-SNE] Iteration 800: error = 0.7239820, gradient norm = 0.0001057 (50 iterations in 3.835s) [t-SNE] Iteration 850: error = 0.6709821, gradient norm = 0.0000933 (50 iterations in 3.907s) [t-SNE] Iteration 900: error = 0.6262846, gradient norm = 0.0000835 (50 iterations in 3.878s) [t-SNE] Iteration 950: error = 0.5880919, gradient norm = 0.0000752 (50 iterations in 3.861s) [t-SNE] Iteration 1000: error = 0.5552707, gradient norm = 0.0000683 (50 iterations in 4.100s) [t-SNE] KL divergence after 1000 iterations: 0.555271 Done.. Creating plot for this t-sne visualization..
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/regression.py:580: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
saving this plot as image in present working directory...
Done performing tsne with perplexity 50 and with 1000 iterations at max [t-SNE] Computing 151 nearest neighbors... [t-SNE] Indexed 19622 samples in 0.059s... [t-SNE] Computed neighbors for 19622 samples in 1.332s... [t-SNE] Computed conditional probabilities for sample 1000 / 19622 [t-SNE] Computed conditional probabilities for sample 2000 / 19622 [t-SNE] Computed conditional probabilities for sample 3000 / 19622 [t-SNE] Computed conditional probabilities for sample 4000 / 19622 [t-SNE] Computed conditional probabilities for sample 5000 / 19622 [t-SNE] Computed conditional probabilities for sample 6000 / 19622 [t-SNE] Computed conditional probabilities for sample 7000 / 19622 [t-SNE] Computed conditional probabilities for sample 8000 / 19622 [t-SNE] Computed conditional probabilities for sample 9000 / 19622 [t-SNE] Computed conditional probabilities for sample 10000 / 19622 [t-SNE] Computed conditional probabilities for sample 11000 / 19622 [t-SNE] Computed conditional probabilities for sample 12000 / 19622 [t-SNE] Computed conditional probabilities for sample 13000 / 19622 [t-SNE] Computed conditional probabilities for sample 14000 / 19622 [t-SNE] Computed conditional probabilities for sample 15000 / 19622 [t-SNE] Computed conditional probabilities for sample 16000 / 19622 [t-SNE] Computed conditional probabilities for sample 17000 / 19622 [t-SNE] Computed conditional probabilities for sample 18000 / 19622 [t-SNE] Computed conditional probabilities for sample 19000 / 19622 [t-SNE] Computed conditional probabilities for sample 19622 / 19622 [t-SNE] Mean sigma: 146.549840 [t-SNE] Computed conditional probabilities in 1.426s [t-SNE] Iteration 50: error = 101.2029419, gradient norm = 0.0048887 (50 iterations in 3.521s) [t-SNE] Iteration 100: error = 77.5842056, gradient norm = 0.0036606 (50 iterations in 4.172s) [t-SNE] Iteration 150: error = 71.2087555, gradient norm = 0.0024406 (50 iterations in 3.771s) [t-SNE] Iteration 200: error = 67.7414703, gradient norm = 0.0018704 (50 iterations in 3.625s) [t-SNE] Iteration 250: error = 65.4947281, gradient norm = 0.0015118 (50 iterations in 3.491s) [t-SNE] KL divergence after 250 iterations with early exaggeration: 65.494728 [t-SNE] Iteration 300: error = 2.5107079, gradient norm = 0.0011589 (50 iterations in 3.498s) [t-SNE] Iteration 350: error = 1.9138236, gradient norm = 0.0006982 (50 iterations in 4.114s) [t-SNE] Iteration 400: error = 1.5128120, gradient norm = 0.0004570 (50 iterations in 3.874s) [t-SNE] Iteration 450: error = 1.2485685, gradient norm = 0.0003241 (50 iterations in 3.643s) [t-SNE] Iteration 500: error = 1.0676084, gradient norm = 0.0002435 (50 iterations in 3.714s) [t-SNE] Iteration 550: error = 0.9377856, gradient norm = 0.0001910 (50 iterations in 3.679s) [t-SNE] Iteration 600: error = 0.8410623, gradient norm = 0.0001555 (50 iterations in 4.224s) [t-SNE] Iteration 650: error = 0.7670552, gradient norm = 0.0001287 (50 iterations in 4.085s) [t-SNE] Iteration 700: error = 0.7090017, gradient norm = 0.0001114 (50 iterations in 4.276s) [t-SNE] Iteration 750: error = 0.6623651, gradient norm = 0.0000969 (50 iterations in 3.696s) [t-SNE] Iteration 800: error = 0.6246436, gradient norm = 0.0000851 (50 iterations in 3.998s) [t-SNE] Iteration 850: error = 0.5937451, gradient norm = 0.0000768 (50 iterations in 3.924s) [t-SNE] Iteration 900: error = 0.5683331, gradient norm = 0.0000701 (50 iterations in 4.342s) [t-SNE] Iteration 950: error = 0.5470531, gradient norm = 0.0000638 (50 iterations in 4.288s) [t-SNE] Iteration 1000: error = 0.5290582, gradient norm = 0.0000600 (50 iterations in 3.969s) [t-SNE] KL divergence after 1000 iterations: 0.529058 Done.. Creating plot for this t-sne visualization..
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/regression.py:580: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
saving this plot as image in present working directory...
Done
train_df = pd.read_csv('pml-training.csv', error_bad_lines=False, index_col=False).drop('Unnamed: 0', axis=1)
test_df = pd.read_csv('pml-testing.csv', error_bad_lines=False, index_col=False).drop('Unnamed: 0', axis=1)
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3146: DtypeWarning: Columns (11,14,19,22,25,70,73,86,87,89,90,94,97,100) have mixed types.Specify dtype option on import or set low_memory=False. has_raised = await self.run_ast_nodes(code_ast.body, cell_name,
train_df = train_df.sample(frac=1).reset_index(drop=True)
train_df.describe()
| raw_timestamp_part_1 | raw_timestamp_part_2 | num_window | roll_belt | pitch_belt | yaw_belt | total_accel_belt | max_roll_belt | max_picth_belt | min_roll_belt | ... | var_yaw_forearm | gyros_forearm_x | gyros_forearm_y | gyros_forearm_z | accel_forearm_x | accel_forearm_y | accel_forearm_z | magnet_forearm_x | magnet_forearm_y | magnet_forearm_z | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 1.962200e+04 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 406.000000 | 406.000000 | 406.000000 | ... | 406.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 | 19622.000000 |
| mean | 1.322827e+09 | 500656.144277 | 430.640047 | 64.407197 | 0.305283 | -11.205061 | 11.312608 | -6.667241 | 12.923645 | -10.436453 | ... | 4639.849068 | 0.157951 | 0.075175 | 0.151245 | -61.651819 | 163.655896 | -55.291917 | -312.575884 | 380.116445 | 393.613745 |
| std | 2.049277e+05 | 288222.879958 | 247.909554 | 62.750255 | 22.351242 | 95.193926 | 7.742309 | 94.594252 | 8.005960 | 93.616774 | ... | 7284.972361 | 0.648618 | 3.100725 | 1.754483 | 180.593687 | 200.130082 | 138.396947 | 346.958482 | 509.373742 | 369.268747 |
| min | 1.322490e+09 | 294.000000 | 1.000000 | -28.900000 | -55.800000 | -180.000000 | 0.000000 | -94.300000 | 3.000000 | -180.000000 | ... | 0.000000 | -22.000000 | -7.020000 | -8.090000 | -498.000000 | -632.000000 | -446.000000 | -1280.000000 | -896.000000 | -973.000000 |
| 25% | 1.322673e+09 | 252912.250000 | 222.000000 | 1.100000 | 1.760000 | -88.300000 | 3.000000 | -88.000000 | 5.000000 | -88.400000 | ... | 0.274550 | -0.220000 | -1.460000 | -0.180000 | -178.000000 | 57.000000 | -182.000000 | -616.000000 | 2.000000 | 191.000000 |
| 50% | 1.322833e+09 | 496380.000000 | 424.000000 | 113.000000 | 5.280000 | -13.000000 | 17.000000 | -5.100000 | 18.000000 | -7.850000 | ... | 612.214225 | 0.050000 | 0.030000 | 0.080000 | -57.000000 | 201.000000 | -39.000000 | -378.000000 | 591.000000 | 511.000000 |
| 75% | 1.323084e+09 | 751890.750000 | 644.000000 | 123.000000 | 14.900000 | 12.900000 | 18.000000 | 18.500000 | 19.000000 | 9.050000 | ... | 7368.414252 | 0.560000 | 1.620000 | 0.490000 | 76.000000 | 312.000000 | 26.000000 | -73.000000 | 737.000000 | 653.000000 |
| max | 1.323095e+09 | 998801.000000 | 864.000000 | 162.000000 | 60.300000 | 179.000000 | 29.000000 | 180.000000 | 30.000000 | 173.000000 | ... | 39009.333330 | 3.970000 | 311.000000 | 231.000000 | 477.000000 | 923.000000 | 291.000000 | 672.000000 | 1480.000000 | 1090.000000 |
8 rows × 122 columns
train_df['classe']=train_df['classe'].astype('category')
Few Pair plots to show corelations between two variables
pp1=train_df.filter(items=['num_window', 'gyros_belt_x', 'gyros_belt_y', 'accel_belt_x', 'accel_belt_y', 'magnet_belt_x','magnet_belt_y', 'classe'])
sns.pairplot(pp1, hue='classe', plot_kws = {'alpha': 0.6, 'edgecolor': 'k'},size = 4)
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/axisgrid.py:1912: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
<seaborn.axisgrid.PairGrid at 0x7f95b30d7ca0>
pp2=train_df.filter(items=['num_window', 'gyros_belt_z', 'accel_belt_z', 'magnet_belt_z', 'classe'])
sns.pairplot(pp2, hue='classe', plot_kws = {'alpha': 0.6, 'edgecolor': 'k'},size = 4)
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/axisgrid.py:1912: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
<seaborn.axisgrid.PairGrid at 0x7f95a8446790>
pp3=train_df.filter(items=['num_window', 'gyros_arm_x', 'gyros_arm_y', 'accel_arm_x', 'accel_arm_y', 'magnet_arm_x','magnet_arm_y', 'classe'])
sns.pairplot(pp3, hue='classe', plot_kws = {'alpha': 0.6, 'edgecolor': 'k'},size = 4)
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/seaborn/axisgrid.py:1912: UserWarning: The `size` parameter has been renamed to `height`; please update your code. warnings.warn(msg, UserWarning)
<seaborn.axisgrid.PairGrid at 0x7f95a909f130>
X=train.drop(['classe'], axis=1)
Y=train['classe']
X_train, X_test,Y_train,Y_test = train_test_split(X, Y, test_size = 0.30, random_state=1)
X_train.shape
(13735, 57)
Y_train.shape
(13735,)
from sklearn.metrics import classification_report
from sklearn.metrics import accuracy_score
from sklearn.svm import SVC
svc=SVC()
svc.fit(X_train, Y_train)
Y_pred_svc=svc.predict(X_test)
print("SVC report \n", classification_report(Y_pred_svc,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_svc,Y_test))
SVC report
precision recall f1-score support
A 1.00 0.29 0.45 5887
B 0.00 0.00 0.00 0
C 0.00 0.00 0.00 0
D 0.00 0.00 0.00 0
E 0.00 0.00 0.00 0
accuracy 0.29 5887
macro avg 0.20 0.06 0.09 5887
weighted avg 1.00 0.29 0.45 5887
Accuracy Score 0.29064039408866993
/Users/vivekyellanti/opt/anaconda3/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1221: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior. _warn_prf(average, modifier, msg_start, len(result))
from sklearn.neighbors import KNeighborsClassifier
KNN=KNeighborsClassifier()
KNN.fit(X_train, Y_train)
Y_pred_KNN=KNN.predict(X_test)
print("K-nearest neighbors Classifier report \n", classification_report(Y_pred_KNN,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_KNN,Y_test))
K-nearest neighbors Classifier report
precision recall f1-score support
A 0.56 0.38 0.45 2536
B 0.21 0.25 0.23 964
C 0.22 0.23 0.22 983
D 0.19 0.22 0.20 802
E 0.16 0.29 0.21 602
accuracy 0.30 5887
macro avg 0.27 0.27 0.26 5887
weighted avg 0.36 0.30 0.32 5887
Accuracy Score 0.30202140309155767
parameters_KNN = {
'n_neighbors': (1,10, 1),
'leaf_size': (20,40,1),
'p': (1,2),
'weights': ('uniform', 'distance'),
'metric': ('minkowski', 'chebyshev')}
clf_knn = GridSearchCV(KNeighborsClassifier(), parameters_KNN,cv=5,n_jobs=-1)
clf_knn.fit(X_train,Y_train)
Y_pred_eknn=clf_knn.predict(X_test)
print("K-nearest neighbors Classifier report \n", classification_report(Y_pred_eknn,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_eknn,Y_test))
K-nearest neighbors Classifier report
precision recall f1-score support
A 0.63 0.61 0.62 1773
B 0.40 0.45 0.42 1006
C 0.46 0.39 0.42 1194
D 0.43 0.38 0.40 1083
E 0.42 0.53 0.46 831
accuracy 0.48 5887
macro avg 0.47 0.47 0.47 5887
weighted avg 0.49 0.48 0.48 5887
Accuracy Score 0.48360794971972143
print(clf_knn.best_params_)
{'leaf_size': 20, 'metric': 'minkowski', 'n_neighbors': 1, 'p': 1, 'weights': 'uniform'}
from sklearn.naive_bayes import GaussianNB
GaussNB=GaussianNB()
GaussNB.fit(X_train, Y_train)
Y_pred_GNB=GaussNB.predict(X_test)
print("Gaussian Naive Bayes report \n", classification_report(Y_pred_GNB,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_GNB,Y_test))
Gaussian Naive Bayes report
precision recall f1-score support
A 0.59 0.71 0.64 1406
B 0.49 0.50 0.50 1118
C 0.54 0.38 0.44 1469
D 0.45 0.41 0.43 1028
E 0.47 0.58 0.52 866
accuracy 0.52 5887
macro avg 0.51 0.52 0.51 5887
weighted avg 0.52 0.52 0.51 5887
Accuracy Score 0.5170715135043316
GaussNB_classifier = GaussianNB()
params_NB = {'var_smoothing': np.logspace(0,-9, num=100)}
Gauss_NB = GridSearchCV(estimator=GaussNB_classifier,
param_grid=params_NB,
cv=3,
verbose=1,
scoring='accuracy')
Gauss_NB.fit(X_train, Y_train)
Gauss_NB.best_params_
Fitting 3 folds for each of 100 candidates, totalling 300 fits
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers. [Parallel(n_jobs=1)]: Done 300 out of 300 | elapsed: 28.6s finished
{'var_smoothing': 4.328761281083061e-09}
Y_pred_GaussNB=Gauss_NB.predict(X_test)
print("Gaussian Naive Bayes report \n", classification_report(Y_pred_GaussNB,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_GaussNB,Y_test))
Gaussian Naive Bayes report
precision recall f1-score support
A 0.61 0.68 0.64 1545
B 0.48 0.53 0.50 1042
C 0.52 0.38 0.44 1412
D 0.45 0.40 0.43 1052
E 0.46 0.58 0.51 836
accuracy 0.52 5887
macro avg 0.50 0.51 0.50 5887
weighted avg 0.52 0.52 0.51 5887
Accuracy Score 0.5163920502802786
print(Gauss_NB.best_params_)
{'var_smoothing': 4.328761281083061e-09}
from sklearn.ensemble import AdaBoostClassifier
AdaBoost=AdaBoostClassifier()
AdaBoost.fit(X_train, Y_train)
Y_pred_AdaBoost=AdaBoost.predict(X_test)
print("AdaBoost Classifier report \n", classification_report(Y_pred_AdaBoost,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_AdaBoost,Y_test))
AdaBoost Classifier report
precision recall f1-score support
A 0.73 0.96 0.83 1297
B 0.53 0.58 0.55 1046
C 0.76 0.50 0.60 1545
D 0.52 0.49 0.51 1021
E 0.77 0.82 0.79 978
accuracy 0.67 5887
macro avg 0.66 0.67 0.66 5887
weighted avg 0.67 0.67 0.66 5887
Accuracy Score 0.6675726176320707
from sklearn.tree import ExtraTreeClassifier
Randtree=ExtraTreeClassifier()
Randtree.fit(X_train, Y_train)
Y_pred_Randtree=Randtree.predict(X_test)
print("Extremely Randomized Trees Classifier report \n", classification_report(Y_pred_Randtree,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_Randtree,Y_test))
Extremely Randomized Trees Classifier report
precision recall f1-score support
A 0.94 0.95 0.95 1683
B 0.87 0.88 0.88 1142
C 0.88 0.86 0.87 1047
D 0.87 0.85 0.86 976
E 0.91 0.92 0.92 1039
accuracy 0.90 5887
macro avg 0.90 0.89 0.90 5887
weighted avg 0.90 0.90 0.90 5887
Accuracy Score 0.9011381009002888
from sklearn.ensemble import BaggingClassifier
BaggingC=BaggingClassifier()
BaggingC.fit(X_train, Y_train)
Y_pred_BaggingC=BaggingC.predict(X_test)
print("Bagging Classifier report \n", classification_report(Y_pred_BaggingC,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_BaggingC,Y_test))
Bagging Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1716
B 0.99 1.00 0.99 1144
C 1.00 0.98 0.99 1036
D 0.99 1.00 0.99 942
E 1.00 1.00 1.00 1049
accuracy 0.99 5887
macro avg 0.99 0.99 0.99 5887
weighted avg 0.99 0.99 0.99 5887
Accuracy Score 0.9947341600135893
BaggingClassifier_parameters = {"n_estimators":list(range(50,60))}
clf_BaggingC = GridSearchCV(BaggingClassifier(), BaggingClassifier_parameters ,cv=3, verbose=1)
clf_BaggingC.fit(X_train, Y_train)
Fitting 3 folds for each of 10 candidates, totalling 30 fits
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers. [Parallel(n_jobs=1)]: Done 30 out of 30 | elapsed: 6.2min finished
GridSearchCV(cv=3, estimator=BaggingClassifier(),
param_grid={'n_estimators': [50, 51, 52, 53, 54, 55, 56, 57, 58,
59]},
verbose=1)
Y_pred_BaggC=clf_BaggingC.predict(X_test)
print("Bagging Classifier report \n", classification_report(Y_pred_BaggC,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_BaggC,Y_test))
Bagging Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1711
B 1.00 1.00 1.00 1148
C 1.00 0.99 0.99 1031
D 0.99 0.99 0.99 950
E 0.99 1.00 1.00 1047
accuracy 1.00 5887
macro avg 1.00 1.00 1.00 5887
weighted avg 1.00 1.00 1.00 5887
Accuracy Score 0.9962629522677086
print(clf_BaggingC.best_params_)
{'n_estimators': 59}
from sklearn.ensemble import GradientBoostingClassifier
GradBC=GradientBoostingClassifier()
GradBC.fit(X_train, Y_train)
Y_pred_GradBC=GradBC.predict(X_test)
print("Gradient Boosting Classifier report \n", classification_report(Y_pred_GradBC,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_GradBC,Y_test))
Gradient Boosting Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1716
B 0.99 0.99 0.99 1148
C 0.99 0.98 0.99 1033
D 0.98 0.98 0.98 958
E 0.98 1.00 0.99 1032
accuracy 0.99 5887
macro avg 0.99 0.99 0.99 5887
weighted avg 0.99 0.99 0.99 5887
Accuracy Score 0.9906573806692712
from sklearn.tree import DecisionTreeClassifier
DTreeC=DecisionTreeClassifier()
DTreeC.fit(X_train, Y_train)
Y_pred_DTreeC=DTreeC.predict(X_test)
print("Decision Tree Classifier report \n", classification_report(Y_pred_DTreeC,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_DTreeC,Y_test))
Decision Tree Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1711
B 0.98 0.99 0.98 1137
C 0.98 0.97 0.98 1033
D 0.98 0.97 0.97 961
E 0.98 0.99 0.99 1045
accuracy 0.98 5887
macro avg 0.98 0.98 0.98 5887
weighted avg 0.98 0.98 0.98 5887
Accuracy Score 0.983692882622728
from sklearn.ensemble import RandomForestClassifier
RandFC=RandomForestClassifier()
RandFC.fit(X_train, Y_train)
Y_pred_RandFC=RandFC.predict(X_test)
print("Random Forest Classifier report \n", classification_report(Y_pred_RandFC,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_RandFC,Y_test))
Random Forest Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1712
B 1.00 1.00 1.00 1152
C 1.00 0.99 0.99 1028
D 0.99 0.99 0.99 948
E 0.99 1.00 1.00 1047
accuracy 1.00 5887
macro avg 1.00 1.00 1.00 5887
weighted avg 1.00 1.00 1.00 5887
Accuracy Score 0.9959232206556821
RandomForestClassifier_parameters = {"n_estimators":list(range(10,20)),"criterion": ["gini", "entropy"]}
clf_RandFC = GridSearchCV(RandomForestClassifier(), RandomForestClassifier_parameters, cv=3, verbose=1)
clf_RandFC.fit(X_train, Y_train)
Fitting 3 folds for each of 20 candidates, totalling 60 fits
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers. [Parallel(n_jobs=1)]: Done 60 out of 60 | elapsed: 30.8s finished
GridSearchCV(cv=3, estimator=RandomForestClassifier(),
param_grid={'criterion': ['gini', 'entropy'],
'n_estimators': [10, 11, 12, 13, 14, 15, 16, 17, 18,
19]},
verbose=1)
Y_pred_RandomFC=clf_RandFC.predict(X_test)
print("Random Forest Classifier report \n", classification_report(Y_pred_RandomFC,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_RandomFC,Y_test))
Random Forest Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1712
B 1.00 1.00 1.00 1150
C 1.00 0.99 0.99 1027
D 0.99 0.99 0.99 950
E 0.99 1.00 1.00 1048
accuracy 1.00 5887
macro avg 1.00 1.00 1.00 5887
weighted avg 1.00 1.00 1.00 5887
Accuracy Score 0.9962629522677086
print(clf_RandFC.best_params_)
{'criterion': 'entropy', 'n_estimators': 19}
Random Forest and Bagging Classifier
BaggC_RandF=BaggingClassifier(RandomForestClassifier())
BaggC_RandF.fit(X_train, Y_train)
Y_pred_BaggC_RandF=BaggC_RandF.predict(X_test)
print("Random Forest and Bagging Classifier report \n", classification_report(Y_pred_BaggC_RandF,Y_test))
Random Forest and Bagging Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1712
B 1.00 0.99 1.00 1153
C 0.99 0.99 0.99 1027
D 0.99 0.99 0.99 951
E 0.99 1.00 0.99 1044
accuracy 1.00 5887
macro avg 0.99 0.99 0.99 5887
weighted avg 1.00 1.00 1.00 5887
print("Accuracy Score",accuracy_score(Y_pred_BaggC_RandF,Y_test))
Accuracy Score 0.995243757431629
Decision Tree and Bagging Classifier
BaggC_DTree = BaggingClassifier(DecisionTreeClassifier())
BaggC_DTree.fit(X_train, Y_train)
Y_pred_BaggC_DTree=BaggC_DTree.predict(X_test)
print("Decision and Bagging Classifier report \n", classification_report(Y_pred_BaggC_DTree,Y_test))
print("Accuracy Score",accuracy_score(Y_pred_BaggC_DTree,Y_test))
Decision and Bagging Classifier report
precision recall f1-score support
A 1.00 1.00 1.00 1716
B 0.99 0.99 0.99 1143
C 1.00 0.98 0.99 1038
D 0.98 0.99 0.99 947
E 0.99 1.00 1.00 1043
accuracy 0.99 5887
macro avg 0.99 0.99 0.99 5887
weighted avg 0.99 0.99 0.99 5887
Accuracy Score 0.9930355019534568
from sklearn.metrics import confusion_matrix
con_matrix_SVC = confusion_matrix(Y_test,Y_pred_svc)
con_matrix_SVC =con_matrix_SVC / con_matrix_SVC.astype(np.float).sum(axis=1)
con_matrix_KNN = confusion_matrix(Y_test,Y_pred_KNN)
con_matrix_KNN =con_matrix_KNN / con_matrix_KNN.astype(np.float).sum(axis=1)
con_matrix_GaussNB = confusion_matrix(Y_test, Y_pred_GaussNB)
con_matrix_GaussNB =con_matrix_GaussNB / con_matrix_GaussNB.astype(np.float).sum(axis=1)
con_matrix_AdaBoost = confusion_matrix(Y_test, Y_pred_AdaBoost)
con_matrix_AdaBoost =con_matrix_AdaBoost / con_matrix_AdaBoost.astype(np.float).sum(axis=1)
f, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, sharex='col', sharey='row', figsize=(10,8))
sns.heatmap(con_matrix_SVC, annot = True, linewidths=.5, ax=ax1, cbar =None, cmap=plt.cm.Blues)
ax1.set_title('SVC')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_KNN, annot = True, linewidths=.5, ax=ax2, cbar =None, cmap=plt.cm.Blues,)
ax2.set_title('KNN')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_GaussNB, annot = True, linewidths=.5, ax=ax3, cbar =None, cmap=plt.cm.Blues)
ax3.set_title('GaussNB')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_AdaBoost, annot = True, linewidths=.5, ax=ax4, cbar =None, cmap=plt.cm.Blues)
ax4.set_title('ABAB')
plt.ylabel('True label')
plt.xlabel('Predicted label')
Text(0.5, 51.00000000000006, 'Predicted label')
from sklearn.metrics import confusion_matrix
con_matrix_GradBC = confusion_matrix(Y_test, Y_pred_GradBC)
con_matrix_GradBC =con_matrix_GradBC / con_matrix_GradBC.astype(np.float).sum(axis=1)
con_matrix_BaggingC = confusion_matrix(Y_test, Y_pred_BaggingC)
con_matrix_BaggingC =con_matrix_BaggingC / con_matrix_BaggingC.astype(np.float).sum(axis=1)
con_matrix_DtreeC = confusion_matrix(Y_pred_DTreeC,Y_test)
con_matrix_DtreeC =con_matrix_DtreeC / con_matrix_DtreeC.astype(np.float).sum(axis=1)
con_matrix_RandFC = confusion_matrix(Y_test, Y_pred_RandFC)
con_matrix_RandFC =con_matrix_RandFC / con_matrix_RandFC.astype(np.float).sum(axis=1)
f, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, sharex='col', sharey='row', figsize=(10,8))
sns.heatmap(con_matrix_GradBC, annot = True, linewidths=.5, ax=ax1, cbar =None, cmap=plt.cm.Blues)
ax1.set_title('Gradient Boosting Classifier')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_BaggingC, annot = True, linewidths=.5, ax=ax2, cbar =None, cmap=plt.cm.Blues,)
ax2.set_title('Bagging Classifier')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_DtreeC, annot = True, linewidths=.5, ax=ax3, cbar =None, cmap=plt.cm.Blues)
ax3.set_title('Decision Tree Classifier')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_RandFC, annot = True, linewidths=.5, ax=ax4, cbar =None, cmap=plt.cm.Blues)
ax4.set_title('Random Forest Classifier')
plt.ylabel('True label')
plt.xlabel('Predicted label')
Text(0.5, 51.00000000000006, 'Predicted label')
print("Extra Randomised tree Classification")
con_matrix_Randtree = confusion_matrix(Y_test, Y_pred_Randtree)
con_matrix_Randtree =con_matrix_Randtree / con_matrix_Randtree.astype(np.float).sum(axis=1)
sns.heatmap(con_matrix_Randtree, annot=True,cmap=plt.cm.Blues)
Extra Randomised tree Classification
<AxesSubplot:>
from sklearn.metrics import confusion_matrix
con_matrix_KNN = confusion_matrix(Y_test, Y_pred_eknn)
con_matrix_KNN =con_matrix_KNN / con_matrix_KNN.astype(np.float).sum(axis=1)
con_matrix_GaussNB = confusion_matrix(Y_test,Y_pred_GaussNB)
con_matrix_GaussNB =con_matrix_GaussNB / con_matrix_GaussNB.astype(np.float).sum(axis=1)
con_matrix_BaggingC = confusion_matrix(Y_test, Y_pred_BaggC)
con_matrix_BaggingC =con_matrix_BaggingC / con_matrix_BaggingC.astype(np.float).sum(axis=1)
con_matrix_RandFC = confusion_matrix(Y_test, Y_pred_RandomFC)
con_matrix_RandFC =con_matrix_RandFC / con_matrix_RandFC.astype(np.float).sum(axis=1)
f, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, sharex='col', sharey='row', figsize=(10,8))
sns.heatmap(con_matrix_KNN, annot = True, linewidths=.5, ax=ax1, cbar =None, cmap=plt.cm.Blues)
ax1.set_title('KNN')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_GaussNB, annot = True, linewidths=.5, ax=ax2, cbar =None, cmap=plt.cm.Blues,)
ax2.set_title('GSNB')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_BaggingC, annot = True, linewidths=.5, ax=ax3, cbar =None, cmap=plt.cm.Blues)
ax3.set_title('Bag C')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_RandFC, annot = True, linewidths=.5, ax=ax4, cbar =None, cmap=plt.cm.Blues)
ax4.set_title('Random Forest Classifier')
plt.ylabel('True label')
plt.xlabel('Predicted label')
Text(0.5, 51.00000000000006, 'Predicted label')
con_matrix_BaggingC_RandFC = confusion_matrix(Y_test, Y_pred_BaggC_RandF)
con_matrix_BaggingC_RandFC=con_matrix_BaggingC_RandFC / con_matrix_BaggingC_RandFC.astype(np.float).sum(axis=1)
con_matrix_BaggingC_DtreeC = confusion_matrix(Y_test, Y_pred_BaggC_DTree)
con_matrix_BaggingC_DtreeC=con_matrix_BaggingC_DtreeC / con_matrix_BaggingC_DtreeC.astype(np.float).sum(axis=1)
f, (ax1, ax2) = plt.subplots(1, 2, sharex='col', sharey='row', figsize=(8,4))
sns.heatmap(con_matrix_BaggingC_RandFC, annot = True, linewidths=.5, ax=ax1, cbar =None, cmap=plt.cm.Blues)
ax1.set_title('Bagging Classifier-RF')
plt.ylabel('True label')
plt.xlabel('Predicted label')
sns.heatmap(con_matrix_BaggingC_DtreeC, annot = True, linewidths=.5, ax=ax2, cbar =None, cmap=plt.cm.Blues)
ax2.set_title('Bagging Classifier-DT')
plt.ylabel('True label')
plt.xlabel('Predicted label')
Text(0.5, 15.0, 'Predicted label')
